Integrated Workflow for Structural Proteomics Studies Based on Cross

Jul 18, 2016 - Integrated Workflow for Structural Proteomics Studies Based on Cross-Linking/Mass Spectrometry with an MS/MS Cleavable Cross-Linker...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

Integrated Workflow for Structural Proteomics Studies Based on Cross-Linking/Mass Spectrometry with an MS/MS Cleavable CrossLinker Christian Arlt,† Michael Götze,‡ Christian H. Ihling,† Christoph Hage,† Mathias Schaf̈ er,§ and Andrea Sinz*,† †

Department of Pharmaceutical Chemistry and Bioanalytics, Institute of Pharmacy, Martin Luther University Halle-Wittenberg, Wolfgang-Langenbeck-St. 4, D-06120 Halle (Saale), Germany ‡ Institute of Biochemistry, Martin Luther University Halle-Wittenberg, Kurt-Mothes-Str. 3, D-06120 Halle (Saale), Germany § Institute for Organic Chemistry, Department of Chemistry, University of Cologne, Greinstr. 4, D-50939 Cologne, Germany S Supporting Information *

ABSTRACT: Cross-linking combined with mass spectrometry (MS) has evolved as an alternative strategy in structural biology for characterizing three-dimensional structures of protein assemblies and for mapping protein−protein interactions. Here, we describe an integrated workflow for an automated identification of cross-linked products that is based on the use of a tandem mass spectrometry (MS/MS) cleavable cross-linker (containing a 1,3-bis-(4oxo-butyl)-urea group, BuUrBu) generating characteristic doublet patterns upon fragmentation. We evaluate different fragmentation methods available on an Orbitrap Fusion mass spectrometer for three proteins and an E. coli cell lysate. An updated version of the dedicated software tool MeroX was employed for a fully automated identification of cross-links. The strength of our cleavable cross-linker is that characteristic patterns of the cross-linker as well as backbone fragments of the connected peptides are already observed at the MS/MS level, eliminating the need for conducting MS3 or sequential CID (collision-induced dissociation)and ETD (electron transfer dissociation)-MS/MS experiments. This makes our strategy applicable to a broad range of mass spectrometers with MS/MS capabilities. For purified proteins and protein complexes, our workflow using CID-MS/MS acquisition performs with high confidence, scoring cross-links at 0.5% false discovery rate (FDR). The cross-links provide structural insights into the intrinsically disordered tetrameric tumor suppressor protein p53. As a timeconsuming manual inspection of cross-linking data is not required, our workflow will pave the way for making the cross-linking/ MS approach a routine technique for structural proteomics studies.

D

features make the chemical cross-linking/MS approach ideally suited for studying intrinsically disordered proteins, which are quite demanding targets for the classical protein structural techniques.6 One of the major drawbacks however is the high complexity of the reaction mixtures created after the cross-linking reaction that pose great challenges for data analysis. There is a need for automated workflows that allow mapping protein−protein interactions on a global level. To date, the use of mixtures composed of light and heavy cross-linkers is the most prominent strategy for a facilitated identification of crosslinked products, and this has paved the way for the analysis of very large protein assemblies up to megadalton sizes and for mapping protein interaction networks.7 Quantitation of crosslinked peptides might be performed using isotope-labeled cross-linkers in case dynamics of protein conformational ensembles are studied8,9 The need for reagents labeled with

uring the last 15 years, chemical cross-linking in combination with a mass spectrometric analysis of the created reaction products has evolved as an alternative strategy for deriving three-dimensional structures of proteins and protein assemblies as well as for mapping protein−protein interactions.1,2 In the chemical cross-linking/mass spectrometry (MS) approach, two functional groups within a protein or between two proteins are covalently connected by a chemical cross-linker. The cross-linker has a defined length and acts as a “molecular ruler” by imposing a distance constraint on the three-dimensional structure of the protein or protein complex. After the cross-linking reaction, the covalently connected amino acids are in most cases identified by liquid chromatography (LC)/tandem mass spectrometry (MS/MS) after enzymatic digestion.3−5 Compared to other methods for protein 3D-structure analysis, the cross-linking/MS strategy is advantageous as sample consumption is low and very large protein assemblies can be investigated. 1 Moreover, in contrast to X-ray crystallography, solution structures of the proteins are obtained, reflecting the overall flexibility of the protein complex. These © 2016 American Chemical Society

Received: December 22, 2015 Accepted: July 18, 2016 Published: July 18, 2016 7930

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

Scheme 1. Chemical Structure of BuUrBu and Theoretical Fragmentation Pathway after Cleavage under Collision-Induced Dissociation Conditions (CID- and HCD-MS/MS)a

Bu: γ-aminobutyric acid; BuUr: isocyanate; P1: peptide 1; P2: peptide 2. Doublet patterns (Δm ∼ 26 u) of reporter ions in the fragment ion mass spectra indicate the presence of a cross-link.

a

In this work, we systematically evaluate different fragmentation methods available on an Orbitrap Fusion mass spectrometer,17 namely, collision-induced dissociation (CID), higher energy collision-induced dissociation (HCD), electron transfer dissociation/collision-induced dissociation (ETciD), and electron transfer dissociation/higher energy collisioninduced dissociation (EThcD), with the aim to optimize the automated identification of cross-linked peptides. Our strategy is validated for cross-linking three proteins, ß-lactoglobulin, bovine serum albumin (BSA), and the tetrameric tumor suppressor protein p53.6 Additionally, we test our workflow for analyzing cross-links in an E. coli cell lysate. For data analysis, an updated version of the in-house software MeroX software is employed, providing the basis for a robust and reliable automated assignment of cross-linked products.18

stable isotopes has resulted in an array of commercially available deuterated cross-linkers.10 MS/MS cleavable cross-linkers, such as disuccinimidyl sulfoxide (DSSO),11 bis(succinimidyl)succinamylaspartyl proline (SuDP),12 bis(succinimidyl)-3-azidomethyl glutarate (BAMG),13 the “Edman-linker”,14 the “urea-linker” (4-{3-[3(2,5-dioxo-pyrrolidine-1-yloxycarbonyl)-propyl]-ureido}-butyric acid 2,5-dioxo-pyrrolidine-1-yl ester, BuUrBu),15 and cyanurbiotindipropionylsuccinimide (CBDPS),16 constitute an alternative highly promising strategy for a reliable automated analysis of cross-linked products. The MS/MS cleavable crosslinker (BuUrBu) used in this work possesses a central urea moiety in its spacer chain that allows one to discriminate different cross-linked species by tandem MS experiments based on their characteristic fragment ion patterns15 (Scheme 1). BuUrBu is compatible with both MALDI (matrix-assisted laser desorption/ionization)- and ESI (electrospray ionization)-MS. The characteristic fragment ion patterns of BuUrBu allow a straightforward identification of different cross-linked species, i.e., intrapeptide (“loop-links”) and interpeptide cross-links as well as modified peptides (“mono-links” or “dead-end” crosslinks) from complex mixtures, and reduce the potential for a false-positive identification of cross-links.

■

EXPERIMENTAL SECTION Material. Full-length wild-type human p53 was produced following an existing protocol.6,19,20 The amine-reactive MS/ MS cleavable urea-based cross-linker (“BuUrBu”) was synthesized by Dr. Francesco Falvo (University of Cologne, Germany) as previously described.15 ß-Lactoglobulin, BSA, and all chemicals were obtained from Sigma-Aldrich (Taufkirchen, Germany). The E. coli cell lysate was prepared 7931

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

Figure 1. Fragment ion mass spectra of exemplary cross-links found in a) p53 and b) E. coli lysate (intramolecular cross-link in pyruvate dehydrogenase E1 component), identified using stepped HCD (most intense precursor ion selection). Spectra were automatically annotated and exported as mgf files by MeroX. 7932

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

variable modification; carbamidomethylation was included as static modification for cysteines. Lys, N-termini, Ser, Thr, and Tyr were considered as potential reaction sites for the BuUrBu; for E. coli data, only Lys and N-termini were considered. Three potential missed cleavage sites were taken into account for each amino acid to be cleaved (Lys and Arg for trypsin; Asp and Glu for AspN).

by sonication, and cell debris was removed by centrifugation. High-performance liquid chromatography (HPLC) solvents were purchased from VWR (Darmstadt, Germany); Milli-Q water was produced by a TKA Pacific system with X-CAD dispenser from Thermo Electron LED GmbH (part of Thermo Fisher Scientific, Niederelbert, Germany). Cross-Linking Reactions. For the cross-linking reactions with BuUrBu, stock solutions were diluted to a final concentration of 10 μM protein. The cross-linking reactions were conducted in 50 mM 2-[4-(2-hydroxyethyl)piperazine-1yl]ethanesulfonic acid (HEPES), pH 7.2, 300 mM NaCl, 10% (v/v) glycerol. For p53, 2.5 mM 2-Tris(carboxyethyl)phosphine (TCEP) was added to the buffer. A freshly prepared stock solution of BuUrBu (in dimethyl sulfoxide, DMSO) was added at 50-fold molar (for p53) or 100-fold molar (for ßlactoglobulin and BSA) excess to the protein solution. E. coli cell lysate was used at a protein concentration of 1 mg/mL; cross-linker was employed at a concentration of 5 mM. The reaction was quenched after 120 min at 4 °C by addition of NH4HCO3 to a final concentration of 20 mM. In-solution digestion was performed with trypsin, while for p53 a mixture of GluC and trypsin was employed.6 Nano-HPLC/Nano-ESI-MS/MS. Peptide mixtures were separated on an Ultimate 3000 RSLC nano system (precolumn: C8, Acclaim PepMap, 300 μm × 5 mm, 5 μm, 100 Å, separation column: C18, Acclaim PepMap, 75 μM × 250 mm, 2 μm, 100 Å, Thermo Fisher Scientific). A gradient from 1% to 35% B was used with a flow rate of 300 nL/min (90 min gradient time; 300 min for E. coli cell lysate), 35% to 85% B (5 min) and 85% B (5 min); solvent A: water + 0.1% formic acid; solvent B: acetonitrile + 0.08% formic acid. The nano-HPLC system was directly coupled to an Orbitrap Fusion Tribrid mass spectrometer with Nanospray Flex Ion Source (Thermo Fisher Scientific). Data were acquired in data-dependent MS/MS mode. Each high-resolution full scan (R = 120 000) in the orbitrap was followed by high-resolution product ion scans (R = 15 000) within 5 s, starting with the most intense signal in the full scan mass spectrum (isolation window of 2 Th); the target value was set to 50 000, and maximum accumulation time was set to 200 ms. Dynamic exclusion (exclusion duration of 60 or 120 s, respectively; exclusion window of ±2 ppm) was enabled to detect less abundant ions. For stepped HCD (30 normalized collision energy (NCE), ±3%) and ETD-based methods, two precursor selection modes were employed: (i) top 5 s/most intense (MI) and (ii) top 5 s/highest charge, then most intense (HC). Charge states >2 were selected for fragmentation. Identification of Cross-Linked Products. Raw data were converted into mgf files using the Proteome Discoverer 2.0 beta (Thermo Fisher Scientific). MS/MS spectra were grouped on the basis of identical precursors (mass tolerance of 2 ppm; elution time window of 90 s). Cross-linked products were analyzed with the in-house software MeroX 1.6.0.18 MS and MS/MS data were automatically analyzed and annotated. MeroX was operated in the reporter ion scan event (RISE) mode where only MS/MS data showing characteristic fragments (two sets of doublets) according to the BuUrBu fragmentation (Scheme 1) were selected and considered for analysis. Results were filtered with false discovery rate (FDR) cutoff of 5%. The detection of intrapeptide cross-links and dead-end cross-links was disabled. A maximum mass deviation of 3 ppm for precursor ions and 10 ppm for fragment ions between calculated and experimental masses was applied as well as a signal-to-noise ratio of ≥2. Oxidation of Met was set as

■

RESULTS The goal of this study was to evaluate different fragmentation methods (CID, HCD, ETciD, and EThcD) available on an Orbitrap Fusion mass spectrometer with the aim of identifying as many cross-linked products as possible within a given chromatographic time frame (90 and 300 min). Also, the different fragmentation techniques were scrutinized for their potential of conducting fully automated cross-link data analyses. ETD alone was not used in this study as the BuUrBu crosslinker did not yield characteristic fragment ions under these conditions. Initial studies had revealed that stepped HCD and EThcD deliver optimum cross-link identification (Figure S1). We therefore focused in subsequent experiments exclusively on these methods to test our automated workflow for the model proteins ß-lactoglobulin, BSA, and the p53 tetramer as well as for an E. coli cell lysate. For an automated data analysis, we first had to modify the MeroX software18 to significantly accelerate the analysis of cross-linked products. Reporter Ion Scan Event (RISE) in MeroX. When MS/ MS cleavable cross-linkers are employed, the characteristic reporter ion patterns created for interpeptide cross-links are the basis for an unambiguous assignment. For BuUrBu used in this study, the MS/MS spectrum is scanned for two matching doublet signals with Δm ∼ 26 u (Scheme 1). Then, the masses of the two corresponding peptides are calculated from the matching doublets, and peptide masses are compared to the in silico digested proteins to identify a cross-link candidate. The quality of MS/MS signal assignment determines the score. Exemplary cross-link identifications with MeroX are presented in Figures 1, S3, and S4. We therefore introduce a new search algorithm in MeroX, termed “reporter ion scan event” (RISE) mode, where exclusively MS/MS data showing the characteristic reporter ions of the MS/MS cleavable cross-linker are selected and considered for analysis, shortening analysis times significantly. As such, total processing times (including in silico digestion, loading, and searching of mgf files) comprise 72 s for a BSA data set with 14 350 MS/MS spectra. When entering a FASTA file containing BSA plus the whole E. coli proteome (4306 entries, Swiss Prot; accession date: 05.05.2016), the total calculation time of MeroX was 31 min 28 s. The resulting cross-linking data set contained 15 544 spectra. Setting the FDR cutoff at 5% yielded 110 cross-links (Figure 2). This underlines the strength of our approach as it is able to discriminate true cross-links in BSA in the presence of the whole E. coli proteome. Calculation times decreased to 4 min 30 s for 2000 proteins and to 2 min 7 s for 1000 proteins. Comparison of Fragmentation Methods. For stepped HCD and EThcD, we first determined the number of MS/MS spectra recorded during a 90 min LC gradient. Not surprisingly, EThcD delivered a lower number of recorded MS/MS spectra compared to HCD (EThcD: 9006 spectra; stepped HCD: 13 482 spectra (average numbers); Figure S2), ultimately resulting in a lower number of potential cross-links, as ETDbased methods are simply more time-consuming than HCD 7933

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

for stepped HCD and 51/23 (most intense) and 45/23 (highest charge) for EThcD. In Figure 4, the numbers of

Figure 2. Screenshot of MeroX decoy analysis with a FASTA file containing BSA and the whole E. coli proteome; the cross-linking data set contained 15 544 spectra. The FDR cutoff was set to 5%, and 110 cross-links met this criterion.

(∼12−16 scan events for EThcD versus ∼20−30 scan events for stepped HCD, using a maximum accumulation time of 200 ms). Evaluation of Cross-Linked Product Assignment. Next, we performed a qualitative comparison of the different fragmentation methods with respect to their ability to unambiguously discriminate true from false-positive crosslinks. The cross-links that had been automatically identified by MeroX were reduced to deliver “unique cross-links” by combining identical cross-linked products exhibiting different charge states as well as different methionine oxidation states. It should be noted in this context that the cross-linking sites identified in BSA and ß-lactoglobulin are in good agreement with published 3D-structures (BSA, pdb entry 3V03; ßlactoglobulin, pdb entry 3NQ9). For p53, the cross-links confirm and extend the findings of our previous structural studies,6 indicating that the structure of the p53 tetramer is more compact than perceived by the existing SAXS data (Figure 3).21,22 Conclusively, the numbers of p53 cross-links (identified/ unique) are as follows: 107/80 (most intense precursor selection) and 106/86 (highest charge precursor selection)

Figure 4. Comparison of identified and unique cross-links detected in (a) ß-lactoglobulin, (b) BSA, and (c) p53. “Identified” cross-links: cross-links detected by MeroX; “unique” cross-links: nonredundant cross-links where different charge states and methionine oxidation states are combined. Precursor ion selection methods: MI: top 5 s/ most intense; HC: top 5s/highest charge, then most intense.

identified cross-links are shown for the three proteins under investigation. A complete summary of all cross-linked peptides is presented in the Supporting Information. The numbers of initially identified cross-links using the RISE algorithm of MeroX are higher for stepped HCD compared to EThcD, irrespective of the precursor ion selection method. Considering the quality of an automated cross-link assignment, stepped HCD and EThcD perform equally well, but given the overall higher number of recorded spectra, stepped HCD outperforms EThcD.

Figure 3. Model of the p53 tetramer based on SAXS data.21 P53 crosslinks identified with our workflow are visualized with xVis.22 The p53 domains are color-coded; TAD: transactivation domain (salmon); DBD: DNA-binding domain (azur); linker (gray); Tet: tetramerization domain (red); Reg: regulatory domain (blue). 7934

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

Figure 5. Venn diagrams showing the overlap of stepped HCD and EThcD in combination with different precursor ion selection methods (MI: most intense; HC: highest charge). The total numbers as well as the percentages of unique cross-links found are given for (a) ß-lactoglobulin, (b) BSA, and (c) p53. The grayscale reflects percentage values.

with the E. coli proteome containing 4306 (Swiss Prot; accession date 05.05.2016) proteins. In addition, we applied longer gradients (300 min) and extended dynamic exclusion times (120 s) to optimize the selection of cross-links (Table S1). 12 634 potential cross-links were found using that optimized workflow with stepped HCD (most intense precursor ion selection), however without FDR cutoff. Using a 90 min gradient, 2965 potential cross-links were identified. For EThcD, the numbers of potential cross-links dropped to 591. Conclusively, as we had already perceived on the basis of the results obtained for single proteins, stepped HCD seems to be the method of choice for identifying cross-links from complex mixtures.

Additionally, we evaluated our results in terms of unique cross-link identification. For BSA (Figure 5a), 19% of unique cross-links were identified with both stepped HCD and EThcD applying different criteria for precursor ion selection. With stepped HCD, a total of 66 unique cross-links were identified; selection of the most intense precursor ion signals (MI) resulted in 14 unique cross-links, while selection of the precursors with the highest charge states (HC) yielded 21 unique cross-links. The overlap was less than half of the crosslinks (31), stressing the importance of precursor ion selection. For the other two proteins under investigation (ß-lactoglobulin and p53), the absolute numbers of cross-links were different, but cross-link identification rates were comparable for the different precursor selection and fragmentation strategies (Figure 5b,c). In general, EThcD delivered lower numbers of unique cross-links. We therefore conclude that stepped HCD should be preferred over EThcD for maximizing cross-link identification. Analysis of E. coli Cell Lysate. To test our workflow for a complex sample, we applied the conditions established on the basis of the three proteins (see above) to an E. coli cell lysate,

■

DISCUSSION Workflow for Single Proteins. The basis of this study is the use of an MS/MS cleavable cross-linkers, such as BuUrBu,15 that allows an automated assignment of cross-links based on characteristic reporter ions. The MeroX software used for data analysis is comfortable to use due to its graphical interface and can handle cross-linking data of medium complexity, i.e., ca. 7935

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

Interestingly, the decoy method used for determining the FDR seems to influence the number of true cross-links (Table S2). For complete proteomes, the numbers of isobaric tryptic peptides comprise on average five or more cross-linked species, requiring almost complete backbone cleavage to unambiguously identify a specific cross-linked product. We did not prefractionate the E. coli cell lysate, which might lead to a coisolation of species. In contrast, the workflow proposed by the Heck group23 employed a time-consuming prefractionation of their HeLa cell lysate. Conclusively, for a thorough and confident assignment of cross-linked products from whole proteomes, preceding fractionation and enrichment steps seem to be indispensable.

4300 proteins can be analyzed in ca. 37 min. The newly introduced RISE mode of MeroX greatly facilitates and accelerates an automated data analysis of cross-links based on characteristic fragment ions created by an MS/MS cleavable linker. Additional advantages of MeroX are that it is not restricted to one single cross-linker and that raw files can be directly converted into mgf files, eliminating the need for additional deisotoping and deconvolution steps. This is the basis for our integrated workflow that does not require timeconsuming data processing prior to cross-link analysis by MeroX. A direct comparison of recorded spectra (Figure S2) favors HCD over EThcD. The faster generation of HCD-MS/MS spectra however leads to a repeated fragmentation of precursors that are eluting over a long time and as such exceed the time window of dynamic exclusion. Therefore, these ions were excluded by a longer dynamic exclusion time (120 s) in order to fragment also low-abundant cross-linked species. The obvious strength of our MS/MS cleavable BuUrBulinker is that the characteristic doublets of the cross-linker as well as backbone fragments of the connected peptides are already observed on the MS/MS level. This eliminates the need for conducting MS3 experiments, which are often of poor quality if low-intensity MS/MS precursors are fragmented. An alternative to MS3 experiments is to combine ETD with HCD in a sequential manner. The resulting spectrum will include high-intensity signals of the characteristic cross-linker fragments (HCD) as well as backbone coverage of the cross-linked peptides (ETD plus HCD, c- and z-type ions supported by band y-type ions). That workflow was recently proposed by the Heck group, allowing the identification of cross-links from endogenous protein complexes in human cellular lysates.23 Their work was based on the application of the MS/MScleavable cross-linker DSSO developed by the Huang group,11 sequential CID- and ETD-MS/MS acquisitions, and the dedicated search engine XlinkX. Our BuUrBu-linker is different from DSSO as the energy required for cleaving the cross-linker is higher for the central urea moiety in BuUrBu compared to the sulfoxide group in DSSO. Thus, we observe both the characteristic doublets of the cross-linker and backbone fragments of the connected peptides on the MS/MS level for BuUrBu. DSSO, on the other hand, shows predominately cross-linker fragments and almost no backbone fragments in MS/MS spectra, requiring MS3 or sequential CID/ETD experiments to delineate the amino acid sequences of the connected peptides. To define differences between the two workflows, we performed a direct comparison of BSA cross-linked with BuUrBu using MeroX and XlinkX, identifying 60 true cross-links with MeroX and 31 with XlinkX (using the linear mode), respectively. For searching our BuUrBu cross-linking data sets with XlinkX in the favored linear mode, we first had to remove precursor signals from the fragment ion spectra and perform deisotoping and deconvolution in order to convert the raw data into mgf files. In summary, XlinkX seems to be optimized for the fragmentation characteristics of DSSO, while MeroX performs better for BuUrBu. Workflow for Complex Samples. In addition to investigating purified proteins, we also performed a cross-link analysis of the whole E. coli proteome with MeroX. When analyzing large data sets, it became obvious that, without inclusion of an FDR, the initial numbers of potential cross-links are very high, but the number of true cross-links is dramatically reduced when the cross-links are scored at 5% FDR.

■

CONCLUSIONS

■

ASSOCIATED CONTENT

On the basis of a thorough comparison of different fragmentation techniques available on an Orbitrap Fusion mass spectrometer, we propose an integrated workflow for a comprehensive, robust, and automated analysis of cross-linked products. The use of MS/MS cleavable cross-linker presents a large step toward a reliable assignment of cross-linked products as specific fragment ions are generated under tandem MS conditions. Considering the per se higher number of identified cross-links in stepped HCD experiments, mainly due to the higher number of MS/MS spectra recorded, this fragmentation technique should be favored over EThcD. The major aim should be to maximize the number of cross-links in one single LC/MS/MS analysis. The obvious strength of our MS/MS cleavable BuUrBu-linker is that the characteristic doublets of the cross-linker as well as backbone fragments of the connected peptides are already observed at the MS/MS level. As MS3 experiments are not required for a cross-link assignment, our workflow is applicable to every lab with access to a highresolution mass spectrometer with tandem MS capabilities, such as Q-TOF or Q-Exactive Orbitrap mass spectrometers. For single proteins and protein complexes our workflow performs with high confidence, scoring cross-links with an FDR above 0.5%. In case whole proteomes are investigated, prefractionation and enrichment steps are required for a confident assignment of cross-links.

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b04853. Evaluation of the quality of automated cross-link assignment by MeroX; comparison of search input for MeroX; MeroX result for cross-linked BSA; MeroX result for cross-linked ß-lactoglobulin; comparison of MeroX calculation times with different E. coli lysate cross-linking data sets; comparison of MeroX identification by applying different decoy methods to identical E. coli lysate cross-linking data sets (PDF) Example files (mgf) for testing MeroX 1.6.0 (ZIP) Summary of cross-linking results for ß-lactoglobulin, BSA, and p53 (xlsx files); summary of MeroX result files (zhrm files) (ZIP) 7936

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937

Article

Analytical Chemistry

■

(21) Tidow, H.; Melero, R.; Mylonas, E.; Freund, S. M.; Grossmann, J. G.; Carazo, J. M.; Svergun, D. I.; Valle, M.; Fersht, A. R. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 12324−12329. (22) Grimm, M.; Zimniak, T.; Kahraman, A.; Herzog, F. Nucleic Acids Res. 2015, 43, W362−W369. (23) Liu, F.; Rijkers, D. T.; Post, H.; Heck, A. J. Nat. Methods 2015, 12, 1179−1184.

AUTHOR INFORMATION

Corresponding Author

*Phone: +49-345-5525170. Fax: +49-345-5527026. E-mail: [email protected]. Notes

The authors declare no competing financial interest. MeroX is available free of charge and can be obtained via the Internet at http://www.stavrox.com.

■

ACKNOWLEDGMENTS The authors declare no competing financial interest. A.S. acknowledges financial support from the DFG (projects Si 867/ 15-1 and 15-2), the EU, and the region of Saxony-Anhalt. M.G. is funded by the DFG (FOR 855 “Cytoplasmic regulation of gene expression” and GRK 1591 “Post-transcriptional control of gene expression-mechanisms and role in pathogenesis”). Dirk Tänzler and Xiaohan Wang are acknowledged for excellent assistance with protein purification and MeroX data analysis. M.G. is indebted to Prof. Elmar Wahle for continuous support.

■

REFERENCES

(1) Sinz, A. Expert Rev. Proteomics 2014, 11, 733−743. (2) Sinz, A. Mass Spectrom. Rev. 2006, 25, 663−682. (3) Kalisman, N.; Adams, C. M.; Levitt, M. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 2884−2889. (4) Leitner, A.; Joachimiak, L. A.; Bracher, A.; Monkemeyer, L.; Walzthoeni, T.; Chen, B.; Pechmann, S.; Holmes, S.; Cong, Y.; Ma, B.; Ludtke, S.; Chiu, W.; Hartl, F. U.; Aebersold, R.; Frydman, J. Structure 2012, 20, 814−825. (5) Rappsilber, J. J. Struct. Biol. 2011, 173, 530−540. (6) Arlt, C.; Ihling, C. H.; Sinz, A. Proteomics 2015, 15, 2746−2755. (7) Herzog, F.; Kahraman, A.; Boehringer, D.; Mak, R.; Bracher, A.; Walzthoeni, T.; Leitner, A.; Beck, M.; Hartl, F. U.; Ban, N.; Malmstrom, L.; Aebersold, R. Science 2012, 337, 1348−1352. (8) Schmidt, C.; Zhou, M.; Marriott, H.; Morgner, N.; Politis, A.; Robinson, C. V. Nat. Commun. 2013, 4, 1985. (9) Fischer, L.; Chen, Z. A.; Rappsilber, J. J. Proteomics 2013, 88, 120−128. (10) Schmidt, A.; Kalkhof, S.; Ihling, C.; Cooper, D. M.; Sinz, A. Eur. Mass Spectrom. 2005, 11, 525−534. (11) Kao, A.; Chiu, C. L.; Vellucci, D.; Yang, Y.; Patel, V. R.; Guan, S.; Randall, A.; Baldi, P.; Rychnovsky, S. D.; Huang, L. Mol. Cell. Proteomics 2011, 10, M110.002212. (12) Soderblom, E. J.; Goshe, M. B. Anal. Chem. 2006, 78, 8059− 8068. (13) Buncherd, H.; Roseboom, W.; de Koning, L. J.; de Koster, C. G.; de Jong, L. J. Proteomics 2014, 108, 65−77. (14) Dreiocker, F.; Muller, M. Q.; Sinz, A.; Schafer, M. J. Mass Spectrom. 2010, 45, 178−189. (15) Muller, M. Q.; Dreiocker, F.; Ihling, C. H.; Schafer, M.; Sinz, A. Anal. Chem. 2010, 82, 6958−6968. (16) Petrotchenko, E. V.; Serpa, J. J.; Borchers, C. H. Mol. Cell. Proteomics 2011, 10, M110.001420. (17) Senko, M. W.; Remes, P. M.; Canterbury, J. D.; Mathur, R.; Song, Q.; Eliuk, S. M.; Mullen, C.; Earley, L.; Hardman, M.; Blethrow, J. D.; Bui, H.; Specht, A.; Lange, O.; Denisov, E.; Makarov, A.; Horning, S.; Zabrouskov, V. Anal. Chem. 2013, 85, 11710−11714. (18) Gotze, M.; Pettelkau, J.; Fritzsche, R.; Ihling, C. H.; Schafer, M.; Sinz, A. J. Am. Soc. Mass Spectrom. 2015, 26, 83−97. (19) Natan, E.; Hirschberg, D.; Morgner, N.; Robinson, C. V.; Fersht, A. R. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 14327−14332. (20) Bullock, A. N.; Henckel, J.; DeDecker, B. S.; Johnson, C. M.; Nikolova, P. V.; Proctor, M. R.; Lane, D. P.; Fersht, A. R. Proc. Natl. Acad. Sci. U. S. A. 1997, 94, 14338−14342. 7937

DOI: 10.1021/acs.analchem.5b04853 Anal. Chem. 2016, 88, 7930−7937